[Top] [All Lists]

Re: [ontolog-forum] IBM Watson & source code

To: doug@xxxxxxxxxx, "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: David Eddy <deddy@xxxxxxxxxxxxx>
Date: Tue, 15 Feb 2011 09:22:53 -0500
Message-id: <2732BE8A-CEAC-4D34-AAF9-6968B21C1270@xxxxxxxxxxxxx>
Doug -    (01)

On 2011-02-15, at 6:58 AM, doug foxvog wrote:    (02)

> Who would want a 100,000 line program
> to be converted with an 80% probability that each line was converted
> correctly?    (03)

Not sure if I injected that "conversion" idea, but that was NOT my  
intent.    (04)

There is already a niche market of tools that convert old languages  
to new.  Such tools can typically automatically convert approximately  
80% of a language, and the remaining 20% needs to be dealt with  
manually.    (05)

What I asked was could Watson help UNDERSTAND the code so I can  
efficiently be guided to where changes should be made.  It is my  
belief that the "typical application" (whatever that is) is written  
in at least 6 different software languages.    (06)

While seemingly attractive on first blush, actual conversion from say  
COBOL to Java, in practice tends to be a less than desirable  
approach.  Too much disruption.  If we assume the expert on the COBOL  
system is no longer available, the junior expert is likely not to be  
facile in Java... so it's easier/safer/cheaper to remain with COBOL.   
BTW... Java is now also approaching legacy status.    (07)

> A system that uses Watson and knows
> what the reserved words are could provide what non-reserved words most
> frequently appear along with the specified word.  However, an  
> algorithmic
> program could easily list every other non-reserved word that is used
> in conjunction with the specified one, sorting them by type of term:
> subroutine, method, variable of different type, etc.  I don't see that
> Watson would be useful for this.    (08)

That's what I thought.    (09)

As you point out, it's the non-reserved words that need to be  
understood.  Non trivial.  Finding them would be easy.  Understanding  
what they mean is the challenge.    (010)

David Eddy
deddy@xxxxxxxxxxxxx    (011)

781-455-0949    (012)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (013)

<Prev in Thread] Current Thread [Next in Thread>